Download on the behaviour of atoms in an electromagnetic wa ve field

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Bremsstrahlung wikipedia , lookup

Coherent states wikipedia , lookup

Particle in a box wikipedia , lookup

Quantum teleportation wikipedia , lookup

Symmetry in quantum mechanics wikipedia , lookup

X-ray fluorescence wikipedia , lookup

Double-slit experiment wikipedia , lookup

Instanton wikipedia , lookup

Atomic orbital wikipedia , lookup

Quantum state wikipedia , lookup

Tight binding wikipedia , lookup

Matter wave wikipedia , lookup

Atom wikipedia , lookup

Electron configuration wikipedia , lookup

Renormalization group wikipedia , lookup

Topological quantum field theory wikipedia , lookup

Quantum electrodynamics wikipedia , lookup

History of quantum field theory wikipedia , lookup

Hidden variable theory wikipedia , lookup

Scalar field theory wikipedia , lookup

T-symmetry wikipedia , lookup

Renormalization wikipedia , lookup

Rutherford backscattering spectrometry wikipedia , lookup

Bohr model wikipedia , lookup

Canonical quantization wikipedia , lookup

Theoretical and experimental justification for the Schrödinger equation wikipedia , lookup

Wave–particle duality wikipedia , lookup

Hydrogen atom wikipedia , lookup

Atomic theory wikipedia , lookup

Transcript
ON THE BEHAVIOUR OF ATOMS IN AN
ELECTROMAGNETIC WAVE FIELD
BY
H. A. KRAMERS
§ 1. Introduction.
As well known, the properties of the atoms cannot be accoun.ted for on the basis of the classical theory of electrons. Still they
exhibit in many respects a great similarity with the properties
which, on the classical theory, systems consisting of small · electrically charged particles would possess. One of the main problems in the modern theory of atoms is therefore to find to what
extent and in what manner the conceptions and the laws of classical electrodynamics can be used in order to establish a description of the actual properties of atoms. Now, an outstanding nonclassical feature of these properties is a certain element of discontinuity which finds an appropriate expression in Bohr's postulate of the ~xistence of discrete stationary states of an atom,
and in the application of the ideas of Planc'k's quantum-theory
to the closer description of the laws governing these states and
the transitions between them. Thus many atomic properties may
be expressed by laws which roughly spoken are a translation into
a »discontinuous language« of laws which, being derived by
means of classical electrodynamics,
are formulated in a »continuous language<<. The object of this lecture is to show in the
case of a special problem of physical int~rest in what way such
a translation can ·be effected, and it will be seen, how a characteristic feature of the mathematical
formulation consists in
the substistutio11 of the »differentials« occurring in the formulre
of the classical theory, by finite »differences«, referring to the
discrete stationary states or the transitions between them.
§ 2. The quantum theory of periodic systems.
Consider an atomic system consisting of one or more electrified particles, and having the property that its motion, when
144
calculated by means of classical mechanics, is always periodic,
independent of the initial positions and velocities of the particles. Let T be the kinetic energy of the system and w the
frequency of the motion, i.e. the number of periods pro sec.,
and let the. quantity J be defined by the time integral of 2T
taken over one period:
(1/ro
Jo
J=
(1)
2 Tdt.
It is easily proved, e.g .. by means of the theorem of least action,
that the energy E of the system can be considered as a function
of J alone, and that the following relation holds:
dE
dJ
(2)
=W.
The Cartesian coordinates of the particles can, considered as a
function of the time be developed in Fourier series of period
1
The same will therefore
hold for the components
of the
U)
electrical moment of the system.
we will therefore have:
~ =
If
~
00
~1: C1:cos (2mwt
0
is one of these components
+ y1:),
(3)
where the summation has to be extended over integer values of
r, and where the amplitudes C1: and phase-constants
Y1: may be
considered as functions of J. According to the classical electron
theory of radiation, the periodically changing electrical moment
( 3) will give rise to the emission of a set of monochromatic electro-magnetic waves, of frequencies w, 2 w, 3 w a.s.o. and of
intensities proportional to the square of the amplitudes characterising the corresponding Fourier components of the electrical
moment. Due to this radiation the energy of the system would
continually decrease.
In the quantum theory of spectra, the behaviour of the periodic system under consideration is described in the following
way. The system possesses a set of stationary states, which may
be characterised by the integer values 0,1,2,3, ... , which a »quantum number« n may take, and the energy En in the nth state
is the same as the energy in a mechanical state of motion for
which the following relation holds:
J =
where h is Planck's
constant.
nh
(4)
The emission of the spectrum of the atom is brought about
by transitions of the atom from one state n' to another state
n" of lower energy; the waves corresponding
to such a transition are monochromatic,
the frequency
v n'
n"
being
given
by
Bohr's frequency. relation:
{5)
This relation
can also be written
n'
v ,, = (n
n
,
~
n
")
En' - -,-1lil
-
in the following
En"
,,
1ln
=
,
(n -
form.
,, A E
AJ
n )
(6)
LI
where .d E and .d J design the difference of the values of E and
J in the two states involved in the transition.
When we corn pare the expression ( 6) with the expression for
one of the frequencies · which, . on the classical theory, would be
involved in the radiation from the system:
V='lW=TdJ
dE
(7)
where -r is an integer, we see at once that Bohr's frequency condition may be considered as the »translation« of a classical formula into a formula which takes account of the discontinuities
of the quantum theory, the quotient of two differentials
being
replaced by the quotient of two differences, referring to a pair
of stationary states.
If we suppose h to converge to zero, and if at the same time
we keep (n'-n") = -r and E constant, the formula (6) will in the
limit be identical with ( 7). But even if h remains finite the
value of ,, expressed by (6) will still tend to coincide asymptotically with that given by ( 7) for -r = n'-n", if we go to the
limit of very great values of n, i.e. if we consider transitions
for which n'-n" is small compared with the values of n' and n"
themselves. Bohr has postulated
that this asymptotic, coincidence of the res~Hs of the quantum theory with those of the
classical theory not only is restricted to the frequencies of the
spectral lines, but also to the intensities with which these lines
appear and to their state of polarisation,
and he considers the
coincidence in question as evidence of a general law or feature,
l1roper to the behaviour of atomic systems, which is called the
146
correspondence principle.
In the terminology of this principle,
the appearance of a spectral line originating from a transition
of the n' th to the n"th state is said to »correspond« to the pre ..
sence of the _harmonic of frequency
(n'-n") w in the motion,
which the system would perform according to the classical electrodynamical laws.
As an example of the application of (4) of (5) and of the
corresponde~e<~ principle we may mention the quantum theory
of the hydrogen spectrum.
In the hydrogen atom the electron
would, according to the classical theory, describe a closed Keplerian orbit round the nucleus as centre. The dependency between
E and J is for this system expressed by:
E '-
-
2n 2 e4 m
J2
(8)
where e and m represent the charge and mass of an electron.
After the introduction of ( 4), the frequencies of the lines of the
hydrogen spectrum are directly obtained by means of ( 5). The
fact that lines are observed corresponding to all possible values
of the difference n'-n" is, according to the correspondence principle, connected with the circumstance,
that in the Fourier resolution of a Keplerian motion harmonics occur corre..;ponding
to any integer value of -r.
The intensity with which a given spectral line appears is
often expressed
Einstein
by means
in 1917, which
of a coefficient
an'
n"
introduced
by
has such a value that an, dt denotes
n"
the probability that, in a time interval dt, an atom in the state
n' shall spontaneously perform a transition to the state n". If
n:,
is
n
the frequency of the corr,esponding spectral line,· the mean
energy emitted per second ·in the form of radiation of this fre-
y
quency by an atom in the state n' will then be given by an' h y n'
n"
.
n"
At present the theory has not yet ' succeeded, however, in
finding the exact mathematical
expressions for the coefficients
an' . . from
n"'
a consideration
of the behaviour
of the atom
ac-
cording to the classical theory we can, however, derive expressions, which by the correspondence-principle
allow to calculate
these coefficients in the limit of high quantum numbers. If we,
especially, consider a system of one degree of freedom in which
an electrical particle can perform oscillations along a straight
line, the electrical moment can, on the classical theory, be represented by a formula of the type (3). According to the classical theory of radiation, the electromagnetic
energy radiated
pro sec. in the form of waves of frequency -r w will be given
by (21r-rw 14 C~/ 3 cs. where c is the velocity of light. The
asymptotical relations for frequency and intensity holding in
the limit of high quantum numbers can then be expressed by:
- n'
v n" N
n'
n'
(9)
(n' - n") ro,
(2n-(n' - n' .') w)' C!'-n"
S
3c
a,,hV,,N
n
n
,
(10)
It will be convenient
for the following to introduce the conception of virtual oscillator, i. e. of an electrical dipole which harmonically oscillates with frequency ,,n' and with such an amn"'
plitudee An' that the classical radiation from it would exactly
n"
correspond with the radiation of that frequency which in the
mean can be ascribed to each atom in the initial state n'. These
quantities An' which we will call the characteristic amplitudes
n"'
of the transitions, will thus in the case of a linear oscillator be
defined by
n'
n'
a,,hv,,=
n
n
and the asymptotical
relation
n'
( 2n Yn"
3
)4
n'
An"
2
(11)
C8
(10) takes the form
(12)
The action ofmonochromatic
light on an atomic system.
In this paragraph we proceed to the problem which forms
the proper subject of _our lecture, viz. the problem of the reaction of an atom when exposed to an electromagnetic wave field.
For the sake of simplicity we will, just as in the end of the
former paragraph, consider an afomic system consisting of an
electrified particle which can perform oscillations along a straight
line, and moreover we will assume that the light waves falling
on the atom are monochromatic and linearly polarised, the electrical vector being parallel to the direction in which the particle
moves. From the analysis of optical phenomena we know, that
§ 3.
148
under the influence of the light, the atom will become a secondary source of radiation; in the language of the classical theory
of radiation it will act as a dipole which oscillates with the same
frequency as that of the incident light. The amplitude of this
oscillation will be proportional to the amplitude of the electrical vector of the incident light, and its phase will directly
depend on the phase of this vector, i.e. the secondary radiation
will be coh,erent with the primary radiation.
If the electrical
light-vector at the position of the atom is presented by
@ = E cos 2nvt
the oscillating dipole»induced«in
~ -=
(13)
the atom can be represented
P cos (2 nvt - 97).
by
(14)
The ratio PIE and the phase difference 97 are the determining·
factors for the dispersion, absorption and scattering of the incident light by a gas consisting of atoms of the kind described.
In fact, if N ~e the nuniber of atoms per ccm., the index of refraction n will be given by
n2 -
'P
. - 4 n N p cos
E .
1-
(15)
The energy' absorbed from the incident beam will be equal to
n v NP £ "sin 97 pro sec., _~orresponding to an effective absorbing
11 P sin 97
cross section equal to 8 n2 - - - pro atom.
The energy
C
E
scattered by the atoms i~ all directions
of frequency
··
v will
finally
·
be equal
in the form of radiation
to
N {2 nv) 4 P 2
pro sec.
3 c3
If v does not lie in the immediate neighbourhood of one of the
absorption lines of the atoms, sin 97 will be very small, and the
problem of the reaction of the atom on the light will primarily
consist in finding which function can represent the values of the
scattering electrical moment P of the atom as a function of the
frequency v of the incident light; P being reckoned positive or
negative according as the value of 97 is very nearly equal to 0,
or to n.
In order to establish the form of the function P, we will first
investigate the influence which, on the classical theory, the light
will have on the motion of the atomic system. Let the electrical
moment of the atom be ~, and let us assume that in the undisturbed motion, ~ is represented by the series:
~o
+ 00 1
.
= R .Ii- - Ci-e2nr-rrot.
(16)
2
-00
The symbol R means, that the real part has to be taken. The
summation has to be extented over positive and negative integer
values of T: zero included. The amplitudes C's are in general
· complex · quantities, satisfying the relation
(17)
C-i-=C-r.
These amplitudes, as well as the frequency w, may be considered
as functions of the quantity J defined by ( 1). Then, writing
~
=
~o
+
(18)
~1
we find, by means of the theory of perturbations,
that when the
atoms is exposed to the periodical field of force (13), and when
second and higher powers of E are neglected,· ~ 1 will be given
by the double infinite series:
... -
~1 -
R -E ~TI)
~oo ~l'~oo { -:r-To
4 -oo
-oo
·
)
.
_
d C-r
C.r--ro
-dJ -(r-T 0)w+v
-Ti--
(
_
)
C d
C-r--ro
} e2ni(v
dJ -(T-T 0)w+v
(19)
I- i- ro)t
From ( 19) we see that the light has induced in the atom not
only a »forced vibration« of frequency v, but also vibrations
the frequencies of which are of the type v T0w. Only in the
case of a harmonic oscillator (all C's zero except C± 1 ) these
»compound« frequencies will not appear.
Consider first that part of (19), which corresponds to a dipole
of frequency v, and for which consequently r 0 = 0. Representing
it by ~, we can .write it in the form
+
~ .=RE
't'
2
]i1
T
!!_
(~~)
d J (rw)2_,_,2
e2nivt.
(20)
Now this expression can be transformed so as to refer to the
actual properties of the atom, by replacing the differential coefficients in an adequate way by quotients of differentials.
In the
case of the formulre (7) and (6) such a transformation
was
brought about by letting the differences on the right hand term
refer to two stationary states, while the left hand term referred
0
.
•
150
to a transition between these two states. Such a thing is not
possible here, however, since the expression for the scattering
electrical momentum .of the atom, · which appears on the left
hand side of (20), necessarily must represent a property of a
given stationary state of the atom. In agreement with this, the
the right hand side
differences appearing when transforming
will refer to two transitions;
the stationary state in question
being the final state for the one and the initial state for the
other. In fact, the asymptotical relations (9) and (12) suggest
directly that we must transform (20) in the following way:
'
where the left member refers to the scattering momentum in the
nth
stationary state. There has been no question of a rigid deduction of (21) from (20), but there are several reasons which
make it possible for us to believe that formula (21) is right.
First of all, it is easily proved, that in .agreement with the correspondenc~ principle, the formulre (21) and (20) become assymptotically identical in the limit of high quantum numbers.
Further the formula (21) contains only quantities which are
directly measurable by experiment, viz. the frequencies of the
light which the atoms may radiate and the quantities A which
govern the intensity of these radiations.
A peculiar consequence follows from the circumstance that
there do not exist stationary states co.rresponding to all positive and negative values of n. If, for instance, only such states
exist for which n=0,1,2 . .. , a great number of terms appeari _ng
in (21) have to be left out. Denoting by the index abs any quantity r~f erring to transitions which the atom in a given state may
perform by absorption of light to a state of greater energy, and
by the index em a quantity referring to a transition to a state
of less energy, (21) may also be given the form:
m
= ~
1"
h
2
{
~
~
A;b,Vaf,s
2
absVab,-V
,
_
~
~
A;~;,,m}·cos 2nv t
,
2
{22)
em v,.m-V
where the summations have to be extented to all the absorption
and all the emission lines of the atom in · the state under consideration.
In the special case where the atom is in its normal
state,
is possible, and no terms exist belonging to the second summation .
inside the brackets. In other words the subtrahends appearing
in the »differences« by which the classical differentials were
replaced, disappear; a feature which illustrates very clearly the
profound difference between the quantum theory of atoms and
the classical theory of electrons. Formally, the formula thus obtained for the scattering momentum in the normal state of an
atom, coincides with the classical formula for ·the scattering from
an atomic system in which the particles can perform harmonic
oscillations of frequencies v ~bs round their positions of equilibrium. It leads immediately to a dispersion formula of the Helmholtz-Ketteler type, which has been confirmed by experiments
on dispersion in gases and vapours, not only in frequency regions far from the frequencies of the absorption lines which
form singular points for the function ~, but also in the neighbourhood of these ·lines, where ~ assumes very large values
giving rise to the so-called anomalous
dispersion.
In the
immediate neighbourhood of an absorption line, formula (22)
must be expected to fail. This is connected with the fact, that
even the classical formula (19) only holds if vis different from
any of the natural frequencies 1:w of the system, since in case
of coincidence the change of the motion of the eJectron due to
the external wave field would, . due to resonance, no longer be
small. The problem of the corrections to be applied to (22) . in
order to represent the scattering also inside the absorption lines
has not yet found a satisfactory solution.
If the atom is in one of its higher states, also terms belonging .
to the second sum inside the brackets of (22) .appear. In the
neighbourhood of the frequency Yem of an emission line, the atoni
will then give rise to an anomalous dispersion of similar kind
as in the case of an absorption line, with the difference that the
sign of ~~ is reversed. This socalled »negative dispersion« is
closely connected · with the prediction made by Einstein, that
. the atom for such a frequency will exhibit a »negative absorption«, i.e. light waves of this frequency, passing through a great
number of atoms in the state under consideration, will increase
in intensity.
Until now we have only considered those terms in the ex-
viz. the state of smallest energy, no emission of ·radiation
152
pression (19) which·correspond to To= 0, i.e. we have investigated
that part of the classical sca~tering which has the same frequency
as the incident light and is coherent with it, and we have
established the analogon of those terms in the quantum theory.
To the other terms in (19), there correspond, however, also well
defined mathematical expressions in the quantum . theory; . these
expressions mean that the atom, under the influence of the light,
will not only scatter radiation of the same frequency, but also
radiation of other frequencies, the intensity of which, however,
is again proportional to the intensity of the incident light. The
classical frequencies of these radiations being of the form Iv+ Tow!,
we are, from the correspondence principle, naturaUy led to expect that in the nth state of the atom these frequencies will be of
Iv + v: _ 1:0 Ior j v + v:+ 1:0./ ,
where· v; should be put equal
to - vb if b should be larger than a. It would lead us too far to
a
give in detail the transformation
of ( 19) which in these cases
leads to an expression of the amplitudes of the scattering momentum of the atom, and the reader is referred to the second
paper cited in the literature reference . .We will only add a few
remarks regarding the finite results.
If To is positive the scattered frequencies are· of the type
-r and their amplitudes will contain terms of the type
0
the form
v+ v:_
E
n+-r n+1:(
1
4 h An - l' An
v" + -r_ v
n
+ v" + 1-r-r+ v) .
n -
(23)
0
where -r represents some whole number. Besides these terms
also other terms of similar structure appear. In order to obtain
these expressions, it is ~ecessary not only to replace the differential coefficients in (19) by the quotients of differences, but also
to replace the factors which contain the amplitudes C in a non
differentiated form by half the sum of two quantities referring
to two definite transitions.
It is interesting in this connection
to remember that also in the mathematical theory of difference
equations, one is lead from the very beginning not only to consider differences
of t~e type
!
renzenrechnung).
of the type_!_ [{ (x
w
[f <x + w)
+ {(x]
+ ro)- f (x)],
but also sums
(Comp. N. E. N~rlund,
Diffe-
If To is negative,
the scattered
v v:_ v- v:-
frequencies
are partly of the
7 =
7o, corresponding
to values of To for
type +
0
which this expression is posmve; partly of the type IV + + 7 o I =
v~ + i- --v, again corresponding
to values of To for which the last
0
expression is positive. In both cases the amplitudes are again
expressed by sums of terms, the structure ·of each of which is
.similar to that of the expression (23).
The generalization of our considerations to systems of more
degrees of freedom, the motion of which is not of a simply periodic, but of a so"'."called multiple periodic type, and the stationary states of which are determined by more than one quantum number, does not involve any difficulty, and no new features
are brought to light by its results. An exception is f ornied,
however, by systems for which the number of quantum numbers
determining the stationary states is less than that of degrees of
freedom. In the case of such systems, which often are called
»degenerate«, and of which the undisturbed hydrogen atom forms
an important example, it has not yet been possible to solve the
problem of the amplitudes of the scattered radiation in a satisfactory way, the difficulty being mainly that of determining the
character of the polarisation of the scattered rays.
The results of which I have tried to give you some idea in this
lecture suggest the possibility of a theoretical treatment of atomic
properties, in which from the outset all formula referring directly
to the classical theory of electrons are banished, and in which
only formulre having a direct meaning in the sense of the quantum theory are allowed. Although it is uncertain on what lines
such a program can be worked out and carried through, interest~
ing and promising results have already been obtained -in the
papers ( 3) and ( 4) mentioned in the literature · list, which
contain an audacious attempt of developing a new »quantum
mechanics«.
v:
LITERATURE
1)
2)
3)
4)
H. A. Kramers:
The quantum theory of dispersion,
Nature 113, 673,
1924; 114, 310, 1924.
H. A. Kramers und W. Heisenberg:
Ober die Streuung von Strahlung
durch Atome, Zs. fiir Phys. 31, 601, 1925.
M. Born:
Ober Quantenmechanik,
Zs. fiir Phys. 26, 379, 1924.
W. Heisenberg: Ober quantentheoretische
Umdeutung kinematischer
und
mechanischer
Beziehungen, Zs. fiir Phys. 33, 879, 1925.